This application represents a continuation of our efforts to utilize mouse models to understand the molecular mechanisms of human prostate cancer, as part of the NCI Mouse Models of Human Cancer Consortium. Our team of investigators includes long-standing members of our previous group (Drs. Cory Abate-Shen and Michael Shen), as well as new members (Drs. Andrea Califano and Carlos Cordon-Cardo) who will bring essential expertise in two key areas, systems/computational biology and systems pathology. At the heart of our program is a series of genetically-engineered mouse models of prostate cancer that have been generated and characterized to identify new biomarkers for biochemical recurrence of human prostate cancer, to characterize key signaling pathways responsible for hormone-refractory prostate cancer, and to pursue pre-clinical studies. In the current application, our objective is to use reverse-engineering of regulatory programs in mouse and human prostate cancer to develop accurate molecular interaction maps for human (human Prostate Cancer interactome, hPCi) and mouse (mPCi). We will employ innovative bioinformatics tools that will identify the "drivers" that are causally responsible for tumor progression, as distinct from other discovery approaches that identify "passenger" genes whose expression is simply correlated with a particular cancer phenotype. Thus, the human and mouse PCis will be interrogated using systems biology tools to elucidate key regulators of the genetic pathways that are dysregulated in cancer (Specific Aim 1) and to identify druggable targets that can affect such altered pathways (Specific Aim 2). Candidate key regulatory genes identified by these analyses will be biochemically and functionally validated for their causal role in tumorigenesis and tumor progression, and their relevance for human prostate cancer will be validated using systems pathology approaches. Finally, lead candidate druggable targets will be evaluated as potential therapeutic targets in pre-clinical trial programs, using our in vivo mouse models.